As well as my technical chops, I’m a bit of an amateur enthusiast when it comes to numbers and statistics. They can be a powerful weapon but equally used badly so I strongly believe in calling out when I see this happening, whether accidentally or otherwise. Recently I’ve had a bit of a Twitter argument with the Police. Or rather a specific Police Officer.

Adam Pace is a recent recruit to Notts Police, have previously been a Special, and uses Twitter, and a special account, to share his experiences and advice. He has over 700 followers, of which I am one of them.

On 10th November, 5pm, he posted a poll on Twitter with a very specific question…

https://twitter.com/CopsInTraining/status/664127827906846722

On 11th November, 7pm – just 26 hours later – he posted the result of the poll…

https://twitter.com/CopsInTraining/status/664523142237433856

And this is where the issue began. I, and others, highlighted to Adam that the wording of the result did not match the reality of the poll posed. “84% of people think that Twitter is a good way for us to keep in better touch with our community” is simply not true. It’s 84% of people on Twitter. More specifically, people on Twitter who were aware of this poll, most of whom are going to be his followers. And with only 26 hours to vote, I’d argue that it was regular Twitter users (and therefore those who like using it) as well. That makes the question biased and when publishing the answer any such bias needs to be made clear.

However, “84% of people on Twitter, who are fans of Twitter, and highlight likely follow this account because they appreciate the information being provided think that Twitter is a good way for us to keep in better touch with our community” doesn’t quite have the same ring to it, does it?

Having said that, bearing in mind how the question was framed, it’s surprising that the result wasn’t higher. Only 84%?

Adam tried arguing with me, eventually resorting to the “it was just an informal poll” reasoning. But is it right? Is it right for someone to misrepresent the results of a poll, informal or not? Is that an acceptable excuse? I’d argue that it isn’t.